AMD Radeon 760M vs NVIDIA RTX PRO 6000 Blackwell Server Comparison

Head-to-Head Benchmarks

The single shared benchmark paints a stark picture, but the context around it is even more revealing. In 3DMark Steel Nomad DX12, the NVIDIA RTX PRO 6000 Blackwell Server records a score of 5996, while the AMD Radeon 760M manages only 400. That is a 93.3% deficit for the AMD part — the NVIDIA card is roughly 15 times faster in this specific test. The delta is so lopsided that it almost feels like a category error, pitting an integrated graphics processor against a server-class accelerator.

Yet the average benchmark scores tell a more nuanced story. The Radeon 760M averages 6019 across all its benchmark entries, while the RTX PRO 6000 Blackwell Server averages 5996. The AMD part actually holds a slight 0.4% edge over the NVIDIA card in overall average, per the nearestRivals data. This is counterintuitive, and it demands explanation. The 760M has ten benchmark results spanning OpenCL, Vulkan, DirectX 9 through 12, and compute workloads; the RTX PRO 6000 has exactly one. That single result happens to be a modern, heavy DX12 load, which is precisely where the NVIDIA architecture is designed to dominate. The AMD part's average is inflated by legacy and compute tests where integrated graphics can hold its own relative to its peer group.

Look at the percentile ranks. The Radeon 760M sits at the 35th percentile of all GPUs, while the RTX PRO 6000 sits at the 34th percentile. These are nearly identical positions, despite the 15x gap in the one shared test. The data suggests that both parts are bracketed by the same rivals: the AMD Radeon RX 6400, NVIDIA GeForce GTX 770M, and NVIDIA Quadro P2000 all appear within 0.5% of the 760M's average, while the RTX PRO 6000's nearest rivals include the same GTX 770M and RX 6400, plus the AMD FirePro W4100 and NVIDIA Quadro K4000M. The server card is not being measured against server parts in this database — it is being measured against a broad swath of consumer and workstation GPUs, and it lands in the same mid-pack neighborhood.

This is the central tension of the comparison: the NVIDIA card wins the only direct head-to-head by a landslide, yet the aggregate statistics place both products at essentially the same level. The explanation lies in benchmark selection, not in raw capability. The 760M's best individual scores — 30,336 in Geekbench Vulkan and 20,255 in Geekbench OpenCL — are respectable for an IGP, but they are not in the same universe as what a 600W Blackwell part should produce. The RTX PRO 6000 simply lacks the breadth of benchmark coverage to show its true position in the aggregate.

Architecture Differences

The architectural chasm between these two parts is the real story. The AMD Radeon 760M is built on a 4 nm TSMC process with 25,390 million transistors packed into a 178 mm² die, yielding a transistor density of 142.6 million per square millimeter. The NVIDIA RTX PRO 6000 Blackwell Server uses a 5 nm TSMC process with 92,200 million transistors on a 750 mm² die, for a density of 122.9 million per square millimeter. The AMD part is denser per square millimeter, but the NVIDIA part has 3.6 times the absolute transistor count and a die more than four times larger. This is the difference between an integrated solution sharing a package with a CPU and a dedicated server accelerator built to fill an entire slot.

The compute resources are not even in the same category. The Radeon 760M has 512 shading units, 32 texture mapping units, 16 ROPs, and 8 ray tracing cores. The RTX PRO 6000 has 24,064 shading units, 752 TMUs, 192 ROPs, and 188 RT cores. The NVIDIA card also brings 752 tensor cores to the table; the AMD part has none listed. The FP32 throughput tells the tale: the 760M delivers 5.323 TFLOPS, while the RTX PRO 6000 delivers 126.0 TFLOPS. Both are rated at a 1:1 FP16 ratio, meaning the NVIDIA card also produces 126.0 TFLOPS in FP16, versus 5.323 for the AMD part. The pixel rate is 41.58 GPixel/s on the 760M versus 502.5 GPixel/s on the RTX PRO 6000. Texture rate is 83.17 GTexel/s versus 1,968.0 GTexel/s.

Memory is where the gap becomes a gulf. The Radeon 760M uses system-shared memory with a system-dependent bus width and bandwidth. The RTX PRO 6000 has 96 GB of GDDR7 on a 512-bit bus, delivering 1.79 TB/s of bandwidth. The AMD part's memory clock is listed as "System Shared," while the NVIDIA card runs at 1750 MHz with 28 Gbps effective speed. There is no scenario in which system-shared memory competes with dedicated GDDR7 at this scale.

The feature sets are similar on paper — both support DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4 — but the execution hardware is vastly different. The RTX PRO 6000 uses a 16-pin power connector and requires a 1000 W suggested PSU, drawing 600 W under load. The Radeon 760M draws 15 W, uses no power connector, and is an IGP with no slot width. The NVIDIA card is dual-slot, measures 267 mm in length, 111 mm in height, and 40 mm in width. The AMD part has no listed dimensions because it is embedded in a processor package.

The Verdict

The data does not support a conventional "which is better" verdict, because the two products are not competitors. The Radeon 760M is an integrated GPU for thin-and-light laptops and mini PCs, with a 15 W TDP and system-shared memory. The RTX PRO 6000 Blackwell Server is a 600 W, dual-slot server accelerator with 96 GB of dedicated GDDR7. The only shared benchmark shows a 93.3% advantage for the NVIDIA part, but the aggregate average scores place both at nearly the same percentile rank, 35th versus 34th. This is a database artifact, not a performance equivalence.

Pick the Radeon 760M if you need a low-power integrated solution for a system that does not have a discrete GPU slot. Pick the RTX PRO 6000 if you need massive compute throughput, ray tracing capability, and server-class memory bandwidth. The data shows that the NVIDIA part wins the only direct comparison by an enormous margin, but the AMD part is the only one of the two that is viable in a mobile or ultra-compact form factor. The 760M's 4 nm process and higher transistor density suggest efficiency-focused design, while the RTX PRO 6000's raw transistor count and die size indicate brute-force performance. There is no overlap in intended use cases.

Specification Differences

| Specification | AMD Radeon 760M | NVIDIA RTX PRO 6000 Blackwell Server |

|---|---|---|

| Process Node | 4 nm | 5 nm |

| Transistors | 25,390 million | 92,200 million |

| Die Size | 178 mm² | 750 mm² |

| Base Clock | 800 MHz | 1590 MHz |

| Boost Clock | 2599 MHz | 2617 MHz |

| Memory | System Shared | 96 GB GDDR7 |

| Memory Bus | System Shared | 512 bit |

| Memory Bandwidth | System Dependent | 1.79 TB/s |

| Shading Units | 512 | 24,064 |

| TMUs | 32 | 752 |

| ROPs | 16 | 192 |

| RT Cores | 8 | 188 |

| Tensor Cores | None | 752 |

| Pixel Rate | 41.58 GPixel/s | 502.5 GPixel/s |

| Texture Rate | 83.17 GTexel/s | 1,968.0 GTexel/s |

| FP32 | 5.323 TFLOPS | 126.0 TFLOPS |

| TDP | 15 W | 600 W |

| Slot Width | IGP | Dual-slot |

| Power Connectors | None | 1x 16-pin |

| Suggested PSU | None | 1000 W |

| Bus Interface | PCIe 4.0 x8 | PCIe 5.0 x16 |

| Display Outputs | Motherboard Dependent | 4x DisplayPort 2.1b |

| Release Date | 2024-01-30 | 2025-03-17 |

| Predecessor | Navi II IGP | Server Hopper |

| Successor | None | Server Rubin |

FAQ

Q: Which GPU has a higher average benchmark score?

A: The AMD Radeon 760M has an average benchmark score of 6019, while the NVIDIA RTX PRO 6000 Blackwell Server averages 5996. The AMD part holds a 0.4% advantage in this metric.

Q: How much faster is the RTX PRO 6000 in the shared 3DMark test?

A: In 3DMark Steel Nomad DX12, the RTX PRO 6000 scores 5996 versus 400 for the Radeon 760M, a delta of 93.3% in favor of the NVIDIA card.

Q: Do both GPUs support the same graphics APIs?

A: Yes, both support DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4.

Q: What is the memory configuration difference?

A: The Radeon 760M uses system-shared memory with system-dependent bandwidth, while the RTX PRO 6000 has 96 GB of GDDR7 on a 512-bit bus with 1.79 TB/s bandwidth.

Q: Which GPU has more ray tracing cores?

A: The RTX PRO 6000 has 188 RT cores, while the Radeon 760M has 8.

Q: What are the TDP ratings?

A: The Radeon 760M is rated at 15 W, while the RTX PRO 6000 is rated at 600 W and requires a 1000 W suggested PSU.

Where Each One Wins

The Radeon 760M wins in efficiency and portability. Its 15 W TDP means it can be integrated into a laptop or compact desktop without additional power delivery infrastructure. The 4 nm process and 142.6M transistors per mm² density indicate a design optimized for power efficiency. Its ten benchmark scores, including a 30,336 in Geekbench Vulkan and 20,255 in Geekbench OpenCL, show it is capable of modest graphics and compute workloads. It wins the aggregate average score comparison at 6019 versus 5996, and it holds a 0.4% edge over the RTX PRO 6000 in that metric.

The RTX PRO 6000 wins in raw performance and capacity. Its 126.0 TFLOPS FP32 throughput, 1.79 TB/s memory bandwidth, and 96 GB of GDDR7 make it suitable for workloads that the Radeon 760M cannot even attempt. The 188 RT cores and 752 tensor cores provide dedicated hardware for ray tracing and AI inference, neither of which the 760M can meaningfully accelerate. The 502.5 GPixel/s pixel rate and 1,968.0 GTexel/s texture rate are orders of magnitude beyond the 760M's 41.58 GPixel/s and 83.17 GTexel/s. It wins the only direct benchmark comparison by a 93.3% margin.

The use-case split is clean: the Radeon 760M is for systems where power draw must stay minimal and the GPU shares memory with the CPU. The RTX PRO 6000 is for server environments where the 600 W power draw, dual-slot footprint, and 1000 W PSU requirement are acceptable trade-offs for massive compute density. The data shows that NVIDIA's part is the only one of the two that can handle modern DX12 workloads at scale, while AMD's part is the only one that can fit into a motherboard's integrated graphics path. Neither product is a substitute for the other, and the benchmark data reflects that reality despite the near-identical aggregate scores.